P

Prime Intellect

Open frontier-model stack for enterprise RL

Member of Technical Staff - Training Platform

Platform EngineeringUpdated on Nov 13
Full-timeSan Francisco, CA, USA

About the Job

Requirements

  • Strong working knowledge of modern AI systems, open model families, fine-tuning techniques such as LoRA, QLoRA, full fine-tuning, RLHF and RLAIF, and inference engines including vLLM, SGLang and TensorRT-LLM.
  • Familiarity with GPU hardware tradeoffs, including H100, H200 and B200 hardware, NVLink, interconnects and memory hierarchy, and their effects on training and inference workloads.
  • Understanding of distributed training fundamentals, including data, tensor, pipeline and expert parallelism, NCCL and multi-node scheduling.
  • Awareness of frontier models, training methods and infrastructure patterns, and ability to translate them into product decisions.
  • Strong Kubernetes operations experience, including Helm, CRDs, operators, KEDA, gang scheduling and GPU operator.
  • Ability to debug production clusters using kubectl and diagnose pod lifecycle, node and networking issues.
  • Cloud platform experience; GCP experience with GCS, GKE, Cloud Run and Cloud Tasks is preferred.
  • Experience with infrastructure automation using Helm, Terraform and Ansible, and a GitOps mindset.
  • Experience with observability tools including Prometheus, Grafana, Loki, OpenTelemetry and DCGM.
  • Linux fundamentals, including networking, namespaces and performance tuning.
  • Strong Python backend development experience, including FastAPI, asynchronous programming and SQLAlchemy.
  • Ability to build Python control-plane agents that interact with Kubernetes APIs.
  • Modern frontend development experience with TypeScript, React or Next.js, Tailwind and shadcn, sufficient to ship product surfaces end to end.
  • REST and tRPC API design experience.
  • Experience building developer tools, dashboards and live-monitoring user interfaces.
  • Experience in either platform or infrastructure development; experience in both is preferred.

Responsibilities

  • Design and operate Kubernetes-based training and inference orchestration across multi-cluster, multi-cloud GPU fleets.
  • Build and maintain Helm charts that compose trainers, inference servers, environment servers and supporting services into reproducible training stacks.
  • Develop Python control-plane agents that watch pods, report run state to the platform and keep clusters in sync.
  • Implement scheduling and autoscaling for heterogeneous hardware using KEDA, LeaderWorkerSet, taints and tolerations, and gang scheduling.
  • Operate a GitOps workflow in which changes ship through pull requests, Helm values and continuous integration.
  • Build node-local model caches, checkpoint pipelines and shared storage to enable fast cold starts.
  • Operate the observability stack and improve the speed of GPU-cluster debugging.
  • Build developer-facing hosted-training features for job submission, live-run monitoring, logs, metrics, model and adapter management, and comparisons.
  • Develop FastAPI backend services and REST APIs that connect the platform to running clusters.
  • Build real-time monitoring and debugging tools for streaming logs, step-level metrics and failure analysis.
  • Ship product user interfaces using Next.js, React, TypeScript, shadcn, Tailwind, tRPC and TanStack Query.
  • Interface with RL trainers, inference servers and environment servers running inside clusters.
  • Productize new training capabilities, including model architectures, RL algorithms and modes.

Desired Qualifications

  • GCP experience, including GCS, GKE, Cloud Run and Cloud Tasks.
  • Experience across both platform and infrastructure development.

About the Company

Prime Intellect

Open frontier-model stack for enterprise RL

Prime Intellect provides an open, end-to-end superintelligence stack for enterprises to train and continually improve their own AI agents. The platform includes scalable compute, large-scale reinforcement learning, environments, sandboxes, evaluations, and deployment tools, and it ships open frontier models to customers. It covers the full lifecycle from training to deployment plus post-training support, enabling agentic training at scale. Its differentiator is the open, end-to-end stack used across thousands of customers with strong backing from top investors, and its goal is to help organizations build and iteratively improve their AI agents using open frontier models.

Company size
51-200
Company stage
Series A
Total funding
$150.5M
Headquarters
San Francisco, CA
Founded
2024

Recently Posted Jobs

Explore more opportunities at Prime Intellect

See more jobs